netdev-dev.d? + udevsend-wait-for-udev?

netdev-dev.d? + udevsend-wait-for-udev?

[Kay Sievers]

I need the net device names in the udev database, for managing the rules
with the gui. Before I send the patch, do we want to raise the dev.d/
events for net devices too?


I have a patch pending for udevsend which waits for the real udev to
come back and let udevsend exit with the return code of the real udev.
It is possible with it, to resubmit the event if we receive a recoverable
error from udev. Don't know if it's useful.
If we want this I will resend a actual version. I want to know it, before
we implement the control socket for the udevd signal handling hole.


thanks,
Kay


-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id\x1470&alloc_id638&op=click
_______________________________________________
Linux-hotplug-devel mailing list  http://linux-hotplug.sourceforge.net
Linux-hotplug-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel

Re: netdev-dev.d? + udevsend-wait-for-udev?

[Greg KH]

On Sat, Mar 27, 2004 at 01:59:28AM +0100, Kay Sievers wrote:
> I need the net device names in the udev database, for managing the rules
> with the gui. Before I send the patch, do we want to raise the dev.d/
> events for net devices too?

Sure, it makes sense to.

> I have a patch pending for udevsend which waits for the real udev to
> come back and let udevsend exit with the return code of the real udev.
> It is possible with it, to resubmit the event if we receive a recoverable
> error from udev. Don't know if it's useful.

I think I lost the last version you sent me of this.  I still don't feel
it's necessary, and possibly dangerous if we have a heavy system load,
right?

But feel free to convince me that we really need this patch :)

thanks,

greg k-h


-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id\x1470&alloc_id638&op=click
_______________________________________________
Linux-hotplug-devel mailing list  http://linux-hotplug.sourceforge.net
Linux-hotplug-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel

Re: netdev-dev.d? + udevsend-wait-for-udev?

[Kay Sievers]

On Fri, Mar 26, 2004 at 05:05:33PM -0800, Greg KH wrote:
> On Sat, Mar 27, 2004 at 01:59:28AM +0100, Kay Sievers wrote:
> > I need the net device names in the udev database, for managing the rules
> > with the gui. Before I send the patch, do we want to raise the dev.d/
> > events for net devices too?
> 
> Sure, it makes sense to.

Fine, then we will have the netdev's in the selection list soon :)

> > I have a patch pending for udevsend which waits for the real udev to
> > come back and let udevsend exit with the return code of the real udev.
> > It is possible with it, to resubmit the event if we receive a recoverable
> > error from udev. Don't know if it's useful.
> 
> I think I lost the last version you sent me of this.  I still don't feel
> it's necessary, and possibly dangerous if we have a heavy system load,
> right?

It's not really dangerous and it doesn't take longer for hundreds of
debug disks.
udevsend will just wait a bit longer than udevs timout for a event and
then quit, also without the "ack message".

Hmm, it's nice style to let the caller know what happened, but are
there any cases you can think of, where we have a error and want to
retry? Hmm, ... :)

thanks,
Kay


-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id\x1470&alloc_id638&op=click
_______________________________________________
Linux-hotplug-devel mailing list  http://linux-hotplug.sourceforge.net
Linux-hotplug-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel

Re: netdev-dev.d? + udevsend-wait-for-udev?

[Greg KH]

On Sat, Mar 27, 2004 at 02:26:15AM +0100, Kay Sievers wrote:
> > > I have a patch pending for udevsend which waits for the real udev to
> > > come back and let udevsend exit with the return code of the real udev.
> > > It is possible with it, to resubmit the event if we receive a recoverable
> > > error from udev. Don't know if it's useful.
> > 
> > I think I lost the last version you sent me of this.  I still don't feel
> > it's necessary, and possibly dangerous if we have a heavy system load,
> > right?
> 
> It's not really dangerous and it doesn't take longer for hundreds of
> debug disks.

I was thinking the problem where we don't get around to servicing the
event in time, and it gets resent.  Wasn't that part of this patch?

If not, nevermind, can you resend it to me?

I'm going to be away this weekend, so you don't have to hurry, I'm not
going to be able to apply it until Monday :)

thanks,

greg k-h


-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id\x1470&alloc_id638&op=click
_______________________________________________
Linux-hotplug-devel mailing list  http://linux-hotplug.sourceforge.net
Linux-hotplug-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel

Re: netdev-dev.d? + udevsend-wait-for-udev?

[Kay Sievers]

On Fri, Mar 26, 2004 at 05:33:06PM -0800, Greg KH wrote:
> On Sat, Mar 27, 2004 at 02:26:15AM +0100, Kay Sievers wrote:
> > > > I have a patch pending for udevsend which waits for the real udev to
> > > > come back and let udevsend exit with the return code of the real udev.
> > > > It is possible with it, to resubmit the event if we receive a recoverable
> > > > error from udev. Don't know if it's useful.
> > > 
> > > I think I lost the last version you sent me of this.  I still don't feel
> > > it's necessary, and possibly dangerous if we have a heavy system load,
> > > right?
> > 
> > It's not really dangerous and it doesn't take longer for hundreds of
> > debug disks.
> 
> I was thinking the problem where we don't get around to servicing the
> event in time, and it gets resent.  Wasn't that part of this patch?
> 
> If not, nevermind, can you resend it to me?

udevsend just waits for udev to come back and sends back the return value
to udevsend. There is no auto-resend or something. The udevsend-caller can
decide what to do with a failing udev call.

> I'm going to be away this weekend, so you don't have to hurry, I'm not
> going to be able to apply it until Monday :)

Oh, I will prepare a whole bunch of patches for you in the meantime :)

Kay


-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id\x1470&alloc_id638&op=click
_______________________________________________
Linux-hotplug-devel mailing list  http://linux-hotplug.sourceforge.net
Linux-hotplug-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel

Re: netdev-dev.d? + udevsend-wait-for-udev?

[Marco d'Itri]

On Mar 27, Greg KH <greg@kroah.com> wrote:

> I think I lost the last version you sent me of this.  I still don't feel
> it's necessary, and possibly dangerous if we have a heavy system load,
> right?
Let's not forget about the effect this could have on boot length as
well.

-- 
ciao, |
Marco | [5384 cou8MmCOsq7dc]


-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id\x1470&alloc_id638&op=click
_______________________________________________
Linux-hotplug-devel mailing list  http://linux-hotplug.sourceforge.net
Linux-hotplug-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel

Re: netdev-dev.d? + udevsend-wait-for-udev?

[Kay Sievers]

On Sat, Mar 27, 2004 at 09:58:51AM +0100, Marco d'Itri wrote:
> On Mar 27, Greg KH <greg@kroah.com> wrote:
> 
> > I think I lost the last version you sent me of this.  I still don't feel
> > it's necessary, and possibly dangerous if we have a heavy system load,
> > right?
> Let's not forget about the effect this could have on boot length as
> well.

What do you mean? Events running in parallel and if udevsend sleeps in
parrallel a bit longer, doesn't mean anything, right? Did you try it?

Kay


-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id\x1470&alloc_id638&op=click
_______________________________________________
Linux-hotplug-devel mailing list  http://linux-hotplug.sourceforge.net
Linux-hotplug-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel

Re: netdev-dev.d? + udevsend-wait-for-udev?

[Marco d'Itri]

On Mar 27, Kay Sievers <kay.sievers@vrfy.org> wrote:

> What do you mean? Events running in parallel and if udevsend sleeps in
> parrallel a bit longer, doesn't mean anything, right? Did you try it?
No... :-)
If udevstart runs everything in parallel then this will obviously not be
a problem.

-- 
ciao, |
Marco | [5386 zaEkbOllPlRng]


-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id\x1470&alloc_id638&op=click
_______________________________________________
Linux-hotplug-devel mailing list  http://linux-hotplug.sourceforge.net
Linux-hotplug-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel